Deployment & Operations
This section covers everything you need to deploy, operate, and maintain DataSynth in production environments.
Deployment Options
DataSynth supports three deployment models, each suited to different operational requirements:
| Method | Best For | Scaling | Complexity |
|---|---|---|---|
| Docker / Compose | Small teams, dev/staging, single-node | Vertical | Low |
| Kubernetes / Helm | Production, multi-tenant, auto-scaling | Horizontal | Medium |
| Bare Metal / SystemD | Regulated environments, air-gapped networks | Vertical | Low |
Architecture at a Glance
DataSynth server exposes two network interfaces:
- REST API on port 3000 – configuration, bulk generation, streaming control, health probes, Prometheus metrics
- gRPC API on port 50051 – high-throughput generation for programmatic clients
Both share an in-process ServerState with atomic counters, so a single process can serve REST, gRPC, and WebSocket clients concurrently.
Operations Guides
| Guide | Description |
|---|---|
| Operational Runbook | Grafana dashboards, alert response, troubleshooting, log analysis |
| Capacity Planning | Sizing model, reference benchmarks, disk and memory estimates |
| Disaster Recovery | Backup procedures, deterministic replay, stateless restart |
Security & API
| Guide | Description |
|---|---|
| API Reference | Endpoints, authentication, rate limiting, WebSocket protocol, error formats |
| Security Hardening | Pre-deployment checklist, TLS/mTLS, secrets, container security, audit logging |
| TLS & Reverse Proxy | Nginx, Envoy, and native TLS configuration |
Quick Decision Tree
- Need auto-scaling or HA? – Use Kubernetes.
- Single server, want observability? – Use Docker Compose with the full stack (Prometheus + Grafana).
- Air-gapped or compliance-restricted? – Use Bare Metal with SystemD.